GitLab

Implement __asan_default_options() for the Chrome OSX binary. This function
sets handle_segv=0 iff the binary is executed with --type=nacl-loader.

BUG=323510
TEST=browser_tests pass under ASan on OSX x64
R=cpu@chromium.org, mseaborn@chromium.org

Review URL: https://codereview.chromium.org/96683003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238328 0039d316-1c4b-4281-b951-d872f2087c98

> Attempt to fix audio wedges by restarting all streams on OSX.
> 
> Introduces two new methods to AudioOutputDispatcher:
> CloseStreamsForWedgeFix() and RestartStreamsForWedgeFix().
> 
> Respectively, each method closes or restarts all active
> streams owned by a given dispatcher.  The process is
> completely transparent to upstream clients.
> 
> A new method on AudioManager, FixWedgedAudio() calls
> CloseStreamsForWedgeFix() for all dispatchers and then
> calls RestartStreamsForWedgeFix() afterward.
> 
> FixWedgedAudio() is called by each AudioOutputController
> when a wedge is detected.  Multiple in flight wedge checks
> are serialized by the audio thread.  The hope is that wedges
> will be fixed before the next WedgeCheck() fires.
> 
> While the methods are available on all platforms, FixWedgedAudio()
> is only wired up on OSX.
> 
> BUG=160920
> TEST=unittest. fake wedge and observe stream recreation.
> R=scherkus@chromium.org
> 
> Review URL: https://codereview.chromium.org/61203008

TBR=dalecurtis@google.com

Review URL: https://codereview.chromium.org/101473002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238327 0039d316-1c4b-4281-b951-d872f2087c98

Introduces two new methods to AudioOutputDispatcher:
CloseStreamsForWedgeFix() and RestartStreamsForWedgeFix().

Respectively, each method closes or restarts all active
streams owned by a given dispatcher.  The process is
completely transparent to upstream clients.

A new method on AudioManager, FixWedgedAudio() calls
CloseStreamsForWedgeFix() for all dispatchers and then
calls RestartStreamsForWedgeFix() afterward.

FixWedgedAudio() is called by each AudioOutputController
when a wedge is detected.  Multiple in flight wedge checks
are serialized by the audio thread.  The hope is that wedges
will be fixed before the next WedgeCheck() fires.

While the methods are available on all platforms, FixWedgedAudio()
is only wired up on OSX.

BUG=160920
TEST=unittest. fake wedge and observe stream recreation.
R=scherkus@chromium.org

Review URL: https://codereview.chromium.org/61203008

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238325 0039d316-1c4b-4281-b951-d872f2087c98

When many readbacks will be occuring in a row, we don't want to have
texture create/delete calls constantly churning through GPU memory.
In particular, this has been a severe performance bottleneck on some
intel drivers.

Instead provide a way for the copy request client to provide a texture
to the compositor that it will use, and then the texture can be reused
for as long as the client wishes.

In the case of tab capture, the client (in RenderWidgetHostViewAura)
wants to use the GLHelper to act on the given texture. Since the
GLHelper lives on the main thread, and the compositor draws on the
compositor thread, we provide the texture from the client through
the main thread CopyOutputRequest API rather than hooking directly
into the compositor thread.

I moved the async readback unit tests out to their own file as
there are now quite a few.

Tests:
LayerTreeHostCopyRequestTestCreatesTexture
LayerTreeHostCopyRequestTestProvideTexture

R=piman@chromium.org
BUG=312497

Review URL: https://codereview.chromium.org/99253004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238324 0039d316-1c4b-4281-b951-d872f2087c98

The Mac app launcher now fades away when it is dismissed, so the
animation that resets the search state can be seen while it's fading
out. To match Win/CrOS, this CL moves the clear to when the window is
closed; at the end of the animation.

Adds a unit test: AppListWindowControllerTest.CloseClearsSearch to
cover.

BUG=321541
TEST=Launch something from a search in the Mac App Launcher. While it's
fading out you shouldn't see the apps grid slide back in.

Review URL: https://codereview.chromium.org/98233002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238323 0039d316-1c4b-4281-b951-d872f2087c98

BUG=324169
TEST=manual

Review URL: https://codereview.chromium.org/98213002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238322 0039d316-1c4b-4281-b951-d872f2087c98

BUG=323263

Review URL: https://codereview.chromium.org/90253002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238321 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238320 0039d316-1c4b-4281-b951-d872f2087c98

BUG=240165
NOTRY=true
R=kinuko@chromium.org, nhiroki@chromium.org

Review URL: https://codereview.chromium.org/98653002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238319 0039d316-1c4b-4281-b951-d872f2087c98

The "reasons" field had been a subset of {NO_SERVICE, NO_NETWORK, NOT_READY},
but it is now used only for signaling errors when it has NOT_READY
without other reasons. This can more simply be represented
by a single "reason" field whose value is set in the order:
NO_SERVICE > NO_NETWORK > NOT_READY, i.e., NOT_READY is set
only when other two states are not set.

BUG=237427
R=benwells@chromium.org, yoshiki@chromium.org

Review URL: https://codereview.chromium.org/98493003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238318 0039d316-1c4b-4281-b951-d872f2087c98

Builds on https://codereview.chromium.org/74653002/

What's left after this:
- Full reassociation on ProcessSyncChanges errors.
- Integration Tests
- Stale Session pruning.

Since GetActiveEntry is now deprecated, switched to use GetVisibleEntry instead per bug 273710.

BUG=98892
R=rlarocque@chromium.org, zea@chromium.org

Review URL: https://codereview.chromium.org/79973002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238317 0039d316-1c4b-4281-b951-d872f2087c98

BUG=260538
TEST=unit_tests
R=kinaba@chromium.org

Review URL: https://codereview.chromium.org/101073002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238316 0039d316-1c4b-4281-b951-d872f2087c98

Disabling following flaky tests on Mac:
WebGLInfoBarTest.ContextLossRaisesInfoBar
WebGLInfoBarTest.ContextLossInfoBarReload

TBR=zmo@chromium.org, danakj@chromium.org
BUG=324555
NOTRY=true

Review URL: https://codereview.chromium.org/100603004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238313 0039d316-1c4b-4281-b951-d872f2087c98

R=enne
BUG=

Review URL: https://codereview.chromium.org/100023003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238312 0039d316-1c4b-4281-b951-d872f2087c98

BUG=325055
TBR=gbillock

Review URL: https://codereview.chromium.org/101123002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238311 0039d316-1c4b-4281-b951-d872f2087c98

Looks like this should be likely fixed too?

BUG=240165
R=tzik@chromium.org

Review URL: https://codereview.chromium.org/100883004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238310 0039d316-1c4b-4281-b951-d872f2087c98

BUG=240165
TEST=unit_tests --gtest_filter=ConflictResolverTest.*
NOTRY=true

Review URL: https://codereview.chromium.org/98053002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238307 0039d316-1c4b-4281-b951-d872f2087c98

BUG=146989
R=joi@chromium.org, satorux@chromium.org, tzik@chromium.org
TBR=brettw@chromium.org

Review URL: https://codereview.chromium.org/96413002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238306 0039d316-1c4b-4281-b951-d872f2087c98

TBR=tonyg@chromium.org
NOTRY=True
BUG=None
TEST=None

Review URL: https://codereview.chromium.org/100603003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238305 0039d316-1c4b-4281-b951-d872f2087c98

The current logic forces a BrowsingInstance swap when going to/from
hosted apps, which breaks postMessage.  Rely on RenderViewHostManager
to do a SiteInstance swap instead.

BUG=123007
TEST=postMessage works after navigating a hosted app popup cross-process.

Review URL: https://codereview.chromium.org/86973002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238304 0039d316-1c4b-4281-b951-d872f2087c98

This will make it easier to do common setup work that requires having the
FileSystemContext, such as setting up Quota checking.

The new GotFileSystemContext method takes a callback to finish opening
each different kind of file system.

This change also renames some callbacks to be a little more consistent.

BUG=194304

Review URL: https://codereview.chromium.org/93403002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238303 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238301 0039d316-1c4b-4281-b951-d872f2087c98

They are timing out.

TBR=tonyg@chromium.org
NOTRY=True
BUG=

Review URL: https://codereview.chromium.org/100653004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238299 0039d316-1c4b-4281-b951-d872f2087c98

Linux is a casualty because we currently don't have a clean system to
differentiate the platforms.

TBR=tonyg@chromium.org
NOTRY=True
BUG=

Review URL: https://codereview.chromium.org/101043002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238298 0039d316-1c4b-4281-b951-d872f2087c98

https://code.google.com/p/chromium/issues/detail?id=314945

Patch from Chris Harrelson <chrishtr@gmail.com>

BUG=314945
..


..


Plumbing for for layout rectangle debug information.

BUG=

Review URL: https://codereview.chromium.org/61883006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238297 0039d316-1c4b-4281-b951-d872f2087c98

- Mark Drive.EntryKind and Drive.FileFormat as obsolete. It's removed in r161277.
- Mark Drive.CacheDBOpenStatus as obsolete. It's removed in r219236.

- Mark Drive.InitialFeedLoadTime obsolete and introduce Drive.DirectoryFeedLoadTime.
  The metric is originally meant to measure the duration until
  we fetch partial result for showing quick UI response to the user.
  In the latest codebase, InitialFeedLoadTime no more is the point
  to trigger such role. Rather, DirectoryFeedLoadTime is.

- Use UMA_HISTOGRAM_LONG_TIMES for Drive.EntireFeedLoadTime and rename it
  to Drive.FullFeedLoadTime.
  It frequently goes beyond 10 seconds upperbound of UMA_HISTOGRAM_TIMES.
  (Despite the comment removed by this CL, I believe the location where
  this stat is taken is correct. The code survived several refactorings
  and at some point in history it was inappropriate...)

BUG=229407
NOTRY=true

Review URL: https://codereview.chromium.org/98253003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238296 0039d316-1c4b-4281-b951-d872f2087c98

It fixes logging and in-process duplication counting in tools/linux/procfs.py
at the same time.

BUG=324194
TEST=None
NOTRY=True

Review URL: https://codereview.chromium.org/98423003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238295 0039d316-1c4b-4281-b951-d872f2087c98

BUG=240165
TEST=manual
R=nhiroki@chromium.org, tzik@chromium.org

Review URL: https://codereview.chromium.org/98633003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238294 0039d316-1c4b-4281-b951-d872f2087c98

- Pass clang-format to make the code more readable, better indentation.
- Better docs.
- Reorganize the code to make more sense of the code flow.
- Use GetImageSkiaNamed() instead of calling ToImageSkia() everywhere.
- append _center prefix to some variables.

BUG=None
TEST=content_shell still works
R=jamescook@chromium.org

Review URL: https://codereview.chromium.org/97633002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238293 0039d316-1c4b-4281-b951-d872f2087c98

This change is part of adding chrome.identity.getAccountAndAuthToken.

The token cache in IdentityAPI and the IdentityMintRequestQueue both had
separate code for tranlating the same API parameters to a lookup key for
their internal maps. They now both use ExtensionTokenKey.

The ExtensionTokenKey includes an account_id field which will allow its
users to support multiple accounts in one profile. All callers currently
use the primary account ID, but will be updated in a future change.

This change also fixes a bug where getAuthToken would return tokens
for the wrong user after switching the account on the profile.

BUG=324874, 322284

Review URL: https://codereview.chromium.org/99173004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238292 0039d316-1c4b-4281-b951-d872f2087c98

Previously, the paddings were not considered in the check if the row was visible or not. This patch fixes this bug.

BUG=324819
TEST=manually tested

Review URL: https://codereview.chromium.org/96823002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238291 0039d316-1c4b-4281-b951-d872f2087c98

Now that packaged apps are not quit when using wrench quit, there is no
way to completely shut down Chrome. While this is normally a good thing
as apps are intended to be separate from Chrome, for developers and
other power users, the ability to completely shut down Chrome can be a
useful function. This also blacklists chrome://quit and chrome://restart
from being saved to the user's session to prevent quit or restart loops.

BUG=317414

Review URL: https://codereview.chromium.org/84893002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238290 0039d316-1c4b-4281-b951-d872f2087c98

If the first touch press happens on a touch-region, but then moves out of that
region, and the second press happens again on the touch-region, then the hit-test
for the second finger would fail because the code used to only look at the position
of the first touch point. Fix that to make sure the position of the new touch
point is used for hit-testing.

BUG=321457
R=jamesr@chromium.org, rbyers@chromium.org
TBR=jochen@chromium.org for content_tests.gypi

Review URL: https://codereview.chromium.org/76283003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238289 0039d316-1c4b-4281-b951-d872f2087c98

BUG=307622
TEST=ImmersiveModeControllerAshTestHostedApp.Layout
Manual Test:
1) Enable "Enable immersive fullscreen for non browser windows" on ChromeOS
2) Install a hosted app (I chose "Cut the Rope")
3) Hit <F4>. Make sure that the app goes fullscreen
4) Hover the mouse at the top of the display and make sure that the window header slides out
5) Hit <F4>. Make sure that the app returns to its formal state.
6) Go to popuptest.com/popuptest1.html in a normal browser window
7) Active one of the popup windows and hit <F4>.
8) Make sure that the popup goes fullscreen
9) Hover the mouse at the top of the display and make sure that the window header slides out
10) Hit <F4>. Make sure that the popup returns to its formal state

Review URL: https://codereview.chromium.org/83343006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238288 0039d316-1c4b-4281-b951-d872f2087c98

Enable revealing the top-of-window views in immersive fullscreen when the mouse is hovered over the top edge of the fullscreen window but the fullscreen window is inactive. This is consistent to how hover works elsewhere on ChromeOS (e.g. the maximize/restore button's help bubble)

BUG=319959
TEST=ImmersiveFullscreenControllerTest.Inactive
R=jamescook
TBR=sky (For change to DragTabToImmersiveBrowserOnSeparateDisplay test)

Review URL: https://codereview.chromium.org/97943002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238287 0039d316-1c4b-4281-b951-d872f2087c98

The files are almost exclusively touched by {hashimoto, hidehiko, kinaba, satorux}
in history, and among those hashimoto@ and I am mainly working on Drive currently.

BUG=none
R=satorux@chromium.org, zelidrag@chromium.org

Review URL: https://codereview.chromium.org/98233004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238286 0039d316-1c4b-4281-b951-d872f2087c98

As of http://crrev.com/237473 we can finally have proper initialization of this code path (before we would have no way to tell the difference between a user who has never initialized and one which initialized to an empty list which was pruned out on write).

Review URL: https://codereview.chromium.org/92173003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238285 0039d316-1c4b-4281-b951-d872f2087c98

This should tell the window its correct size, instead of the size expanded to be at least 64 pixels on each side.

BUG=286609,320974

Review URL: https://codereview.chromium.org/87383002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238284 0039d316-1c4b-4281-b951-d872f2087c98

This changes the ownership model of infobars so that InfoBars are long-lived and
own their InfoBarDelegates directly.  The InfoBarService pseudo-owns the
InfoBars (instead of deleting them directly, it tells them when they're unowned
and expects them to delete themselves).

This fixes leaks when infobars are closed while not visible (e.g. in a
background tab) and in general makes the system clearer and easier to reason
about.

BUG=62154
TEST=none
R=erg@chromium.org, sail@chromium.org, sky@chromium.org

Review URL: https://codereview.chromium.org/22694006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238283 0039d316-1c4b-4281-b951-d872f2087c98

These IPCs correspond to Javascript IME API events
(oncandidatewindow*) defined in
https://dvcs.w3.org/hg/ime-api/raw-file/8c061ee19f99/Overview.html#inputmethodcontext-interface

They will fire when IME candidate window changes its appearance.

The Blink side of the change is already checked in:
https://src.chromium.org/viewvc/blink?revision=161139&view=revision

This is the Chromium part of sending IPCs to renderer and
a test for both Chromium and Blink code path, which is not
covered by blink r161139.

The whole change FYI is worked on at
https://codereview.chromium.org/80583002/
and other CLs (to generate actual events from IMEs etc.)
will follow.

BUG=238585
TEST=content_browsertests --gtest_filter="RenderViewImplTest.SendCandidateWindowEvents"

Review URL: https://codereview.chromium.org/84963003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@238282 0039d316-1c4b-4281-b951-d872f2087c98